HCTGS v38 Australia — The Continent That Already Decided

Abstract HCTGS v38 Australia — The Continent That Already Decided. The premise. Australia settled the question of manufacturing water from the sea twenty years ago and built six large membrane plants to do it. This document does not reopen that decision. It addresses the question that follows: what a second generation of coastal water plants should also produce, and what the water should cost once eleven other products carry the structure. The architecture is thermal throughout — separation occurs at the phase change, with no membrane stage and no concentrate returned to the sea. The ceiling and the floor. Australia consumed 17.2 billion cubic metres of water across industry and households in 2023–24, and averaged 15.3 billion over the preceding decade. At 365 million cubic metres per tower-year that is 42 to 47 towers for everything the continent uses. Agriculture alone is 32; the whole of South Australia is 3.3; every domestic tap in the country is 5.1. Hydroelectric throughput is excluded as non-consumptive. Both figures are ceilings rather than plans, and nobody has to accept the roof to act at the floor. The comparison that carries the scale. The largest desalination asset any Australian state has funded cost $1.83 billion and is rated at 100 million cubic metres a year — 0.27 of one tower. Across its realised operating history since 2011, roughly 239 million cubic metres cumulative, it has delivered 0.044 to 0.049 of a tower, running a minimum-production regime for nine months of a typical year and standing down entirely in others. That is correct behaviour for the asset class rather than failure: a single-product plant whose largest operating line is purchased electricity has every reason to stand down when water is cheap. The entire national seawater intake for desalination, 170 million cubic metres a year across six plants, is less than half what one tower draws. The capital line, which a desalination business case conventionally omits, is set out against a Gulf plant of matching output rather than against a lightly used one. That plant produces 1,036,000 cubic metres a day — within four per cent of a tower at reference output — and cost USD 7.2 billion, which bought the desalination trains and the 2,400 MW power station that drives them, with sixteen heavy fuel oil tanks of twenty thousand cubic metres each on site. A tower at the same water output costs USD 0.65 to USD 1.15 billion at series rate, carries 1,000 to 1,500 MW of its own, and holds no fuel, its fuel being the 1,290 tonnes of magnesium arriving dissolved at the intake each day. Per cubic metre a day of installed capacity that is about USD 6,950 against USD 650 to USD 1,150, and over a twenty-five year life with water carrying all of the capital, USD 0.76 per cubic metre against 7 to 13 cents. Three qualifications are attached: the arithmetic is undiscounted capital alone with operating cost left out on both sides; the Saudi plant sells electricity as well as water, so the USD 7.2 billion buys both; and a well-run modern membrane plant on cheap power reaches capital costs per cubic metre close to a tower's, so capital is not where this is decided. What is decided is the operating line — one plant buying fuel and electricity for as long as it stands, the other separating its fuel from the feed it was already processing and selling the ash as oxide. One structure at reference output. One million cubic metres a day, 365 million a year. Electricity 24–36 GWh a day, or hydrogen at 480–720 tonnes a day, being the same capacity spent differently and never additional to it. Magnesia 1,500–1,930 tonnes a day and salt 27,000 tonnes a day, both as capacity. Arriving dissolved at the intake each day: magnesium 1,290 t, calcium 411 t, potassium 392 t, bromine 67.3 t, boron 4.5 t, lithium 170 kg. Nothing is mined; the elements arrive because the plant makes water. For scale inside the country: the whole Snowy Mountains Scheme averages 4,500 GWh a year and holds something in the order of 250 square kilometres of drowned valley behind sixteen dams, while one tower carries two to three times that annual energy on 0.08 to 0.20 square kilometres. Because only vapour makes the phase change, hardness, chloride, fluoride, boron, nitrate, arsenic, microplastics, PFAS, bacteria and viruses remain on the concentrate side, and minerals are dosed back to a specified profile. Two temperature histories, and why the water is cheap. The thermal cascade descends in eight stages from specialist ceramic at 2,200 °C through the magnesium flame at 1,500–1,800 °C, the melt near 700 °C, the magnesium–chlorine circuit at 450–550 °C, pyrohydrolysis from 220 °C, process steam at 110 °C and a 60 °C bulk circuit, down to ambient. The tower column is separate: it begins near 1,500 °C at the ceramic pipe and the hydrogen–oxygen burner, holds near 110 °C over its full height, and ends at a cold core roughly ten metres across at 4 °C in the crown, where a curtain of chilled mist between the cupola wall and the steam performs the whole of the condensation. No desalination plant heats its water through, because that places the full latent load — some 26 gigawatts at reference throughput — onto the fuel bill. This one does not either: at 0.2 bar the brine boils at 60 °C unaided, the vortex separates, and the burner is an igniter and momentum source rather than the bulk heat supply. The suction evaporates; the flame trims and turns the column. The operating intake is the top five to ten centimetres of the sea, and deep water is drawn only for district and compute cooling. What is new, stated narrowly. Every stage of this chemistry has been run somewhere and none of them has been run in series. Pyrohydrolysis of magnesium chloride, sequential crystallisation, calcination and chloride electrolysis are each established industrial practice; what has never been built is the arrangement in which each stage is heated by the stage above it. Run separately, each carries its own fuel bill, and that rather than any difficulty in the separations is why element-by-element recovery from seawater has never paid. The claim is not that the cascade invents a separation but that it stops paying for the same joule eight times. Vortex prior art is named in full — Yen US 4,070,131, Hsu US 4,452,562, Yeh US 4,935,639, Jacobs 1983, Rechenberg 1984, Dick 1983, Mamulashvili 1984–1997, Craven 1998, Rennó and Bluestein 2001 — including the disclosure that forecloses the easy demarcation and therefore narrows the claim to two statements: tangential flame entry below the waterline, and condensation in flight with mineral separation and water production in the same vessel. The concentrate line. A membrane plant is a separator with one saleable output and purchased electricity as its largest operating line; a cascade is a refinery with twelve outputs and electricity as a product. Membranes sort by molecular size and charge against an osmotic gradient and cannot separate one salt from another; a cascade sorts by solubility against a temperature gradient. The two therefore meet at one point only, which is the concentrate. Australia's six plants return most of their intake through outfalls already sited, licensed and instrumented, carrying roughly twice the magnesium per cubic metre of raw seawater. What that line changes is the feed and the site — a richer feed, a smaller column, a site through approval once — with the physics unchanged. The cluster. Ten towers form one installation, sharing not only an intake, outfall, grid connection and perimeter but the cascade itself: one magnesium–chlorine circuit, one pyrohydrolysis train, one mineral separation plant, one electrolysis hall supplying stoichiometric oxygen to every burner in the group. Ten columns feed one ladder, which carries most of the capital, so a later tower approaches the cost of the column alone. The marine footprint is assessed once for all ten. The build is staged: two or three towers first, the remainder against demand, with later towers drawing binder-grade magnesia and salt and gypsum fractions from the cluster's own output. Five coasts, and the line inland. Five Australian coasts are identified as holding industry, a mine within reach and relief behind the shore at the same time: the upper Spencer Gulf, the Illawarra, Gladstone, Townsville and Bunbury–Kemerton. The upper Spencer Gulf is proposed as the first installation — an inverse estuary reported to raise salinity landward, a warm surface layer, the country's only integrated steelworks with its concentrator on site, and its own magnetite ore body 56 km inland. Where coastal relief reaches five to six hundred metres the cistern stands above nearly everything inland, and the command area below it is suppliable without a pump; the head basin is a through-basin with outflow rated at 150 per cent of inflow, intermediate cisterns carry storage duty covering days while the operating window covers the season, and Australia's long, straight, legally reserved inland highway corridors are proposed as the alignment. What the ground needs. Australian soils have weathered in place for tens of millions of years, are among the most phosphorus-deficient on earth, are widely sodic in the subsoil and store cyclic salt in the regolith, so deserts that green readily elsewhere on young loess will not green here on water alone. The country has answered this twice already, on the Western Australian wheatbelt and through mid-century trace-element work on written-off sandplain, which makes the deficiency an input cost rather than a barrier. Calcium, magnesium, potassium and boron come off the ladder and the water is remineralised to a specified profile; phosphorus and nitrogen do not and remain bought inputs. The governing hazard is hydrological: applying water to country that stores salt raises the water table, so drainage and application rate belong in the first design. The industry that forms at the plant